You find the significantTerms streaming expressions useful: https://lucene.apache.org/solr/guide/7_6/stream-source-reference.html#significantterms
Joel Bernstein http://joelsolr.blogspot.com/ On Mon, Jan 21, 2019 at 3:02 PM Pratik Patel <pra...@semandex.net> wrote: > Aman, > > Thanks for the reply! > > I have tried with corrected query but it doesn't solve the problem. also, > my tags filter matches multiple documents, however the interestingTerms > seems to correspond to just the first document. > Here is an example of a query which matches 1900 documents. > > > http://localhost:8081/solr/collection1/mlt?debugQuery=on&q=tags:voltage&mlt.boost=true&mlt.fl=my_field&mlt.interestingTerms=details&mlt.mindf=1&mlt.mintf=2&mlt.minwl=3&q=*:*&rows=100&start=0 > > > Thanks, > Pratik > > > On Mon, Jan 21, 2019 at 2:52 PM Aman Tandon <amantandon...@gmail.com> > wrote: > > > I see two rows params, looks like which will be overwritten by rows=2, > and > > then your tags filter is resulting only one document. Please remove extra > > rows and try. > > > > On Mon, Jan 21, 2019, 08:44 Pratik Patel <pra...@semandex.net wrote: > > > > > Hi Everyone! > > > > > > I am trying to use MLT request handler. My query matches more than one > > > documents but the response always seems to pick up the first document > and > > > interestingTerms also seems to be corresponding to that single document > > > only. > > > > > > What I am expecting is that if my query matches multiple documents then > > the > > > InterestingTerms handler result also corresponds to that set of > documents > > > and not the first document. > > > > > > Following is my query, > > > > > > > > > > > > http://localhost:8081/solr/collection1/mlt?debugQuery=on&fq=tags:test&mlt.boost=true&mlt.fl=mlt.fl=textpropertymlt&mlt.interestingTerms=details&mlt.mindf=1&mlt.mintf=2&mlt.minwl=3&q=*:*&rows=100&rows=2&start=0 > > > > > > Ultimately, my goal is to get interesting terms corresponding to this > > whole > > > set of documents. I don't need similar documents as such. If not with > > mlt, > > > is there any other way I can achieve this? that is, given a query > > matching > > > set of documents, find interestingTerms for that set of documents based > > on > > > tf-idf? > > > > > > Thanks! > > > Pratik > > > > > >
